Development of the nitrided layer mechanism of the growth, mathematical modeling and simulation'

1
(No Transcript)
2
Development of the nitrided layer mechanism of
the growth, mathematical modeling and simulation.

• Jerzy Ratajski, Adam Mazurkiewicz, Dariusz
  Lipinski, Jerzy Dobrodziej
• Koszalin University of Technology, Koszalin,
  Poland
• Institute of Sustainable Technology, Radom,
  Poland

3

• Nitriding process is very efficient
• in long-run production,
• in the mass production,
• and also is very often used in so called duplex
  process

4
Duplex processes
5

• The nitrided layer should characterizes
• demanded hardness and thickness,
• The nitrided parts should maintain their sizes
  and shape stability.

6
(No Transcript)
7
g
a-Fe(N)
e
8
Equilibrium diagram Fe-N
9

• Metods of artificial inteligence
• Neural network
• Fuzzy logic
• Evolutional algorithms

10
Calculation by iteration methods
i 1, 2, 3 1 e 2 g 3 a
etc
11
(No Transcript)
12
Structure of nitrided layer
Structure of nitrided layer
13
(No Transcript)
14
Structure of nitrided layer Hardness distribution
Hp
g600
g500
g400
15
Steel 4340 (AISI) KN 3,25 , T 5800C , t 10 h
16
t 3 h
t 10 h
Steel 4340 KN 3,25 T 5800C
17
Steel 4340
KN 10 T 5800C t 16 h
KN 10/0.5 T 5800C t 8/8 h
18
Conclusion

• A research results presented, indicate that,
  besides temperature and nitrogen potential, also
  phase composition of iron (carbo)nitrides layer
  on steel has considerable influence on
  development of hardness profiles in diffusion
  layer.

19

• Metods of artificial inteligence
• Neural network
• Fuzzy logic
• Evolutional algorithms

20
(No Transcript)
21
Table Characteristics of experimental data used
for modeling.
22

• The established neural model enables
  estimation of influence of the chemical
  composition of steel (grade of steel) and
  nitriding process parameters including number
  of stages, on hardness profile.

23
Summary

• A research results presented show, that also
  phase composition of iron (carbo) nitrides layer
  on steel has considerable influence on
  development of hardness profile in diffusion
  layer.
• Elaborated neural network model constitute tool
  to the simulation of the profiles of hardness in
  the nitrided layer - predicted results showed
  relatively low scatter with experimental results.
  
• The model is open for constant upgrade and
  improvement and also can be applied in a control
  system and in visualization of the process course.

24

• Thank you